Competitors
Competing TeamsSeptember 16-17, 2017 - Conference Hall, Hokkaido University, Sapporo, Japan Track 1 - Smartphone-basedTeam Name Description The positioning technique proposed by our team consists of 2 general stages of database establishment and positioning algorithm. The database establishment stage utilizes information acquired from indoor Wi-Fi network and magnetic field strength acquired from geomagnetic sensor in a complex manner. The positioning algorithm stage consists of initial position determination and pedestrian positioning. Initial position determination uses a Wi-Fi signal-based fingerprint algorithm. Pedestrian positioning then provides position value obtained by compensating initial position value through relative positioning, by means of smartphone’s geomagnetic sensor, accelerometer sensor and direction sensor. This process involves fingerprint positioning using a feature extraction algorithm embodied by applying BoW(Bag of Words) technique.
Team Name Description The real-time SNU-NESL PDR is based on Pedestrian Dead Reckoning (PDR) using built-in inertial sensors and pressure sensor in smartphone for indoor navigation. In the SNU-NESL PDR, step and its length are estimated using an accelerometer, and attitude reference system enables us to estimate the orientation of the device using gyroscope and accelerometer. In order to improve the heading accuracy, the SNU-NESL PDR uses a dominant direction of the building as a measurement. We propose multi-virtual track dominant direction models for heading estimation. Lastly, the floor is determined with the help of pressure difference combined with 2D PDR system.
Team Name Description HaLo is an indoor positioning system of pedestrian dead reckoning (PDR), with WiFi fingerprint and map-matching techniques and is implemented on a smartphone. The proposed method aims to enhance the tracking performance of the whole system with the WiFi fingerprint technique as well as can reduce the building cost of the radio map with a fewer number of reference positions compared to conventional systems. In addition, the methods to detect turning behavior and collisions based on a given map information are suggested to correct the position from the PDR system. The proposed method could reduce the positioning errors by 51.02% in compared to the original PDR method and 37.18% with the PDR aided WiFi fingerprint. Team Name Description In many engineering applications, indoor positioning is required where the Global Positioning System(GPS) is not available. A smartphone-based hand-held indoor positioning system is presented in this paper. The system collects data using the accelerometers, gyroscopes, barometers and gravity sensors embedded in the smartphones. The accelerometer and gravity data are used for zero-velocity detection and calculating the vertical displacement of each walking step, and then the inverted pendulum model is applied to calculate the step length of every step. The angle of direction is estimated by processing gyroscope data with the quaternion method. The step length and the direction angle of each step are combined to determine the coordinates of each step. The barometer is used for measuring the height information. A Kalman filter is used in zero-velocity-update (ZUPT) to reduce the vertical speed offset caused by accelerometer drift errors. Wifi is also fused in our system. In order to guarantee the accuracy, map information and magnetic field information are used in the navigation systems. The experiment results show that we obtained high precision results with common hand-held smartphone often seen on market. Track 2 - Pedestrian Dead Reckoning
Team Name AOE Corresponding Author Wenchao Zhang Affiliation University of Chinese Academy of Sciences, Beijing, China Academy of Opto-electronics, Chinese Academy of Sciences, Beijing, China Description In this plan, a method based on IMU/EKF+HMM+ZUPT+ZARU+HDR+Compass is designed to realize foot-mounted pedestrian navigation. The general range ratio test (GLRT) and the Hidden Markov Model (HMM) were used to realize the detection of zero speed interval at different speed states. When the zero speed state is detected, the zero velocity update (ZUPT) method is used to limit the accumulation of IMU. The Zero Angular Rate Update (ZARU) + (heuristic heading reduction) HDR+Compass method is used to limit the IMU attitude and heading drift. Finally, the EKF method is used to realize the effective estimation and feedback of the speed, attitude and heading error of the pedestrian navigation system. Meanwhile, a fault detection algorithm based on the innovation vector is added to the EKF system to effectively detect and eliminate the gross errors in the measurements. Team Name Description Kalman-based Inertial Navigation System (INS) is a reliable and efficient method to estimate the position of a pedestrian indoors. Classical INS-based methodology which is called IEZ (INS-EKF-ZUPT) makes use of an Extended Kalman Filter (EKF), a Zero velocity UPdaTing (ZUPT) to calculate the position and attitude of a person. However, heading error which is a key factor of the whole Pedestrian Dead Reckoning (PDR) system is unobservable for IEZ-based PDR system. To solve the problem, Electronic Compass (EC) algorithm is a valid method. But magnetic disturbance may lead to heading error. In this paper, the Quasi-static Magnetic field Detection (QMD) method is proposed to detect the pure magnetic field and then selects EC algorithm or Heuristic Drift Reduction algorithm (HDR), which implements the complementation of the two methods. Meanwhile, the QMD, EC and HDR algorithm are integrated into IEZ framework to form a new PDR solution which is named as Advanced IEZ (AIEZ).
Team Name Description For infrastructure-free pedestrian navigation techniques in order to restrict the accumulating sensor error one usually requires a map or building layout of the respective indoor or underground areas. Using the FootSLAM algorithm developed at DLR one can “learn” a suitable map of walkable areas when persons, equipped with foot-mounted inertial sensors, walk in indoor or underground environments, and visit regions repeatedly. The processing algorithm uses machine learning techniques based on Bayesian inference. The FootSLAM algorithm consists of two cascaded filters: a lower Unscented Kalman Filter (UKF) that estimates the odometry of the pedestrian’s walk and an upper SLAM algorithm that reduces the drift and estimates simultaneously the map of the walkable areas of the environment. The FootSLAM map is built upon hexagon prisms and the weighting is based on transition counts of the hexagon faces visited during the walk. Team Name Description The Magneto Inertial Navigation exploits magnetic disturbances usually found in office buildings to reconstruct the device velocity. The presented magneto-inertial dead-reckoning navigation system is equipped with MEMS sensors including an IMU and a magnetometer array. It computes the velocity - and then the position - in a Kalman filter fusing magnetic and inertial information, without relying on hypothesis about the nature of the movement, nor the magnetic field structure. Team Name Description In this paper, a strides detection algorithm is proposed using inertial sensors worn on the ankle. This innovative approach based on geometric patterns can detect both normal walking strides and atypical strides such as small steps, side steps and backward walking that existing methods struggle to detect. It is also robust in critical situations, when for example the wearer is sitting and moving the ankle, while most algorithms in the literature would wrongly detect strides. Team Name Description High accuracy in indoor navigation with foot-mounted sensors attracts a lot of researches in the last decades. This paper present a foot-mounted inertial navigation system. The system can achieve 3D positioning in a variety of gait. Team Name Description We proposed a PDR system using foot-mounted IMU. The data from IMU is processed in the laptop. First, we use gait tracking to estimate the trajectory of the pedestrian. In the gait tracking module, step detection is utilized to calibrate the velocity to get a better estimation. Then, we use map matching algorithm to refine the estimated trajectory and correct the heading direction. With refined trajectory, our system outputs current position. In our preliminary experiments, the error between ground truth position and the estimated one was about 3 meters after a 400 meters walking.
Team Name Description The inertial pocket localization system requires the sensor to be introduced in the front trousers’ pocket. The proposed localization system is purely based on inertial sensors, i.e. accelerometers and gyroscopes. We use the step and heading algorithm, which consists of two main blocks: the computation of the heading and the computation of the displacement. Along with the orientation angles, we also estimate the biases of the gyroscopes. The displacement estimation occurs every time a new step is detected. The location in the pocket allows identifying stairs and this way the vertical displacement is also computed. We use a landmark-based drift compensation algorithm that detects seamlessly stairs and corners while the user walks. These landmarks are associated, when re-visited, and the drift error accumulated over the trajectory is computed. This value is used by the orientation estimation filter that generates a low-drifted heading estimation, which leads to drift-compensated trajectories. Team Name Description Our participating system hardware consists of two parts. The first part, the foot module. The foot module consists of a gyroscope for positioning, a three-axis accelerometer, a barometric altimeter, a magnetometer, a microprocessor chip, and a Bluetooth module for data transmission. The foot PDR module in the actual positioning, is fixed on the foot of the pedestrian, used to calculate pedestrian position, and then the results will be sent out through the Bluetooth. The second part is a general intelligent mobile phone. The mobile phone uses the relative position data sent by the foot PDR module received by Bluetooth, the map data stored on the mobile phone and other related information, runs the positioning fusion, gives the final positioning result and real-time display on the mobile terminal. Track 3 - Smartphone-based (off-site)Team Name Team Name Team Name
Team Name Team Name Description The WiFi based localization method used by the University of Technology Sydney (UTS) team during the competition could be broadly categorized under Wi-Fi Fingerprinting. In contrast to these traditional distance based Fingerprinting approach, we chose a Bayesian framework to fuse information from Wi-Fi signal strength readings to readings from other sensors such as Barometer, Gyroscope and Magnetic sensor that are available in smart phones and a motion model based on the prior knowledge of the floor plan and associated expected nominal motions
Team Name Description With the increasing need of location-based services, indoor localization based on fingerprinting has become an important technique due to its high accuracy and low hardware requirement. In the offline stage, we use a novel algorithm for indoor moving trajectory estimation using multiple sensors to calculate each step between two ground-truth position. After the calculation, we are able to find more coordinates between two ground-truth position. According to the information of AppTimestamp, we then can map these coordinates on to wifi information. In the online stage, with the previous result, we utilize the wifi information of its SSID and RSS value to locate the position. We also use the information of pressure, wifi and GPS to determine the floor difference and building difference. Track 4 - PDR for warehouse picking (off-site)Nagoya UniversityKDDI Research, Inc. (Japan) ETRI (South Korea) Xiamen University (China) Yuan Ze University (Taiwan) |